Analytical Verification of Crack Propagation through Cold Expansion Residual Stress Fields

Transcription

1 Analytical Verification of Crack Propagation through Cold Expansion Residual Stress Fields Aircraft Airworthiness and Sustainment Conference Grapevine, Texas, USA March 22, 2016 Sharon Mellings, John Baynham - CM BEASY Keith Hitchman, Joy Ransom - FTI 1

2 OUTLINE Analysis of engineered residual stress fields using ABAQUS Using residual stress fields in BEASY crack growth analysis Overview of the selected test case BEASY approaches used to match the test case Conclusions and Future Work 2

3 Analysis of engineered residual stress fields using ABAQUS Non-linearities evaluated: True Stress, psi Material non-linearity (plasticity) Geometric non-linearity Displacement Contact Correlation/verification: Physical Test (deformation, strain) Laboratory Test (XRD, contour) Closed-form solutions Stress (psi) Typical material response True Strain, in/in Distance from Edge of Hole (in.) X-Ray Diffraction X-Ray Diff. Hoop 2D FEA Hoop 3D FEA Entry Hoop 3D FEA Exit Hoop X-Ray Diff. Radial 2D FEA Radial 3D FEA Entry Radial 3D FEA Exit Radial 3

4 Analysis of engineered residual stress fields using ABAQUS Split Sleeve Cx Problem Definition Pull mandrel through hole Allow for stress relaxation Final Ream EXPLICIT IMPLICIT Resize ream partition Yes Precise ream? No / Done Remote Load Step Generate Reports Residual Stress (RS) Database Fatigue/Fracture Analysis 4

5 Max Net Section Stress, ksi Analysis of engineered residual stress fields using ABAQUS Residual Stress (RS) Database Integrating Residual Stress Analysis of Critical Fastener Holes into USAF Depot Maintenance Document: A3G ; See Section 4.6 (pp43-48). Developed by APES, ESRD RS Database visualizer Fatigue/Fracture Analysis 2D analyses AFGROW (using RS database) NASGRO 3D discretized analyses XFEM (ABAQUS) FRANC3D STRESSCHECK BEASY SIF Solutions Cycles to Failure Estimates NASGRO and AFGROW predictions 60 AFGROW, No-FtCx NASGRO, No-FtCx 50 AFGROW, FtCx NASGRO, FtCx Cycles 5

6 Analysis of engineered residual stress fields using ABAQUS FEA Examples Composite Structure Analysis Installation Process Optimization Landing Gear Rib with ForceMate High Load Transfer Specimen with Split Sleeve Cold Expansion (section view) 6

7 Using residual stress fields in BEASY crack growth analysis Engineered residual stresses change the crack growth behavior Compressive residual stress = slower growth Tensile residual stress = faster growth Crack growth rates may vary across the crack front Cracks will grow through regions with varying residual stress levels BEASY with FEA results can be used to evaluate these effects. 7

8 Using residual stress fields in BEASY crack growth analysis The process for creating this residual stress simulation using BEASY Create Boundary Element model of the cracked part Add a crack or cracks to the Boundary Element model Apply the residual stress field from FEA to the crack surface Compute stress intensity factors via summation rule Crack growth determined by desired relationship (NASGRO equation, Paris, etc) K K A B K K max( K max( K A K max min B K K R R,0),0) da dn C K m 8

9 FEA Model Geometry Using residual stress fields in BEASY crack growth analysis BEASY BE Model Automated internal step Automated external step FEA Model Res. Stresses Loading Condition Add crack(s) Compute SIFs Propagate Crack(s) Create new crack geometry Update geo? Yes Continue growth? No Yes Generate Reports 9

10 Overview of the selected test case = OH fatigue coupon geometry 5/16 fastener hole 2.35 plate width 0.25 plate thickness 2024-T3 material ys = ksi Paris relationship C = 3.09x10-9 ksi-in ½ m = ksi gross stress@ R=0.05 Load Load Sleeve Gap Element size: X = Y = Z =

11 Stress, psi 20,000 Overview of the selected test case Residual Hoop Stress 10, ,000-20,000-30,000-40,000-50,000 Shaded area shows approximate extent of crack propagation in BEASY Mandrel Entry Midplane Mandrel Exit -60, Radial Distance from hole, inch 11

12 Overview of the selected test case FS Stopped after 37,000 cycles FS Stopped after 17,000 cycles 12

13 BEASY approaches used to match the test case Initial crack geometry (first trials): Entry crack initial radius = The exit crack initial radius = Cracks normal to load (Mode I only) Loading and propagation relationship: Constant amplitude for Mode I 30 ksi maximum gross stress, R = 0.5 Paris relationship, no retardation effects Static residual stress field Simulation Aims: Can we simulate the observed damage? How do the fatigue results compare? Entry crack Exit crack 13

14 BEASY approaches used to match the test case Under the crack growth simulation, the computed SIF results are used to advance the crack New meshes are created and new SIF values computed during the simulation Due to combination of the compressive residual stresses and the dynamic opening load the entry crack grows faster than the exit crack Entry size Numbers are cycle counts to grow the crack (crack nucleation is at cycles) Exit size 14

15 BEASY approaches used to match the test case The fatigue life for this simulation provides generally good agreement with the test results that were obtained This simulation only includes cracks on one side of the hole The simulation was not continued after the cracks joined FS Stopped after 37,000 cycles 15

16 BEASY approaches used to match the test case The simulation with two cracks provides a good agreement with the test results but does not explain some of the crack shapes seen during test The compressive residual stress slows the growth of cracks and therefore it is likely that further new cracks will initiate along the bore during the load cycling This has been evaluated using a third crack along the bore Entry crack New mid crack Exit crack 16

17 BEASY approaches used to match the test case This simulation shows some agreement with crack profiles that have been viewed during the testing This indicates that these profiles could be created by multiple crack nucleation along the bore Numbers are cycle counts to grow the cracks 17

18 BEASY approaches used to match the test case During the life, it is therefore conceivable for many cracks to nucleate along the bore here 5 cracks have been simulated Cycle counts to grow the crack The grown cracks here resembles those seen practice The crack growth was stopped at coalescence With crack geometry update the analysis could be continued to a point resembling the actual crack front. FS Stopped after 37,000 cycles 18

19 Conclusions The simulation work carried out indicates that the crack growth profiles observed in testing cannot be explained purely by the Paris-driven growth of corner cracks initiating in the part Observed crack growth profiles require multiple interacting cracks initiating on the hole bore. Total damage is a combination of the crack initiation and growth. In this scenario, we hypothesize that the initiation of new cracks is more prevalent due to the reduction in the crack growth rate. The simulation should ideally be continued with crack growth of the coalesced cracks until the critical crack size has been reached. 19

20 Future Work Further work could be carried out using this combination of residual stress evaluation and crack growth simulation: Continue current analysis with joined cracks Evaluate with alternate crack growth relationship Looking further ahead to different cases using BEASY: Cracks under different load conditions (CA vs spectrum, compressive loading) Cracks from filled holes Cracks from ForceMated (bushed) holes 20

21 Questions? 21